CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   FLUENT (http://www.cfd-online.com/Forums/fluent/)
-   -   Using turbulence model for steady state analysis (http://www.cfd-online.com/Forums/fluent/28250-using-turbulence-model-steady-state-analysis.html)

Terry February 26, 2001 18:10

Using turbulence model for steady state analysis
 
Hi!

It is understood that we can use turbulence model for Navier-stokes equations at high reynolds number (Spalart-Allmaras). Why turbulence models are needed for steady-state analysis of high reynolds number flows and which features of these flows do they attempt to capture?.

Can anyone help on this issue?.

Thanks


John C. Chien February 27, 2001 04:08

Re: Using turbulence model for steady state analys
 
(1). The need to use a turbulence model comes from the fact that the governing equations for turbulent flow are the Reynolds equations, which contain statistical terms called Reynolds stresses terms and these terms are unknowns and must be modelled first before the equations can be solved. (2). You need to study fluid mechanics for turbulent flows first, in order to get the feeling about the difficulties with turbulent flows. (3). The Reynolds equations are derived from the Navier-Stokes equations about one hundred years ago.

Terry February 27, 2001 07:48

Re: Using turbulence model for steady state analys
 
Hi!

Thanks John.

In the investigation of a flow over a symmetrical aerofoil,then would the use of turbulence model like the spalart allaramas model produce better results as compare to the other models like K-eplison or the laminar model?.

Would the change in the computational grid type and density affect the use of the model?

Thanks.

regards


John C. Chien February 27, 2001 13:52

Re: Using turbulence model for steady state analys
 
(1). For flow over airfoil, you need to know the flow conditions such as the Reynolds number, the Mach number first. This will tell you whether the flow is laminar or not, whether the density is important or not. For civil aircraft flying at transonic speed, the Mach number and density is important. And the flow is mostly turbulent.(2). The grid issue has been discussed here over two years now. You will have to obtain the so-called "grid independent solution" to make sure that the solution does not change with the particular grid used. This means grid refinement. (3). The selection of turbulence model and the development of an improved model are part of research activities to validate the result. You change the model until you are satisfied with the computed results, and in most cases, this is done in comparison with test results. (4). Simple flow over airfoil does not give simple solution, it is rather complicated and you have to consider the flow transition, transonic shock formation, and the shock/boundary layer interaction or possible flow separation, possible flow oscillations. (5). I do not think that the current commercial CFD codes can handle all these problems. You can't have it , if it is not there in the code. But you have something to play with.

Terry February 27, 2001 19:02

Re: Using turbulence model for steady state analys
 
Hi!

It is understand that the Spalart - Allmaras models solves the navier-stokes equations at high reynolds number to achieve the solution for an symmetric aerofoil.

I am using a tri elements and pave type. Is this a suitable form for meshing and predicting the life and drag of a aerofoil close to the ground. ie: ground effect. Is that any way of improving the grid accuracy?.

If I use the Euler equations of inviscid flow or potential flow theory would I be able to solve the solution, if it does? would it be faster?.

Thanks.

Regards

John C. Chien February 27, 2001 23:11

Re: Using turbulence model for steady state analys
 
(1). For ground effect, using Euler or potential flow equation should give you faster results. In this case, it will be inviscid solution. (2). I have been using two-equation models for 25 years, and I don't think the model you mentioned is any better than existing models. I can't comment on that because I don't use it at all. (3). It has been known over 15 years that the triangular mesh is not suitable for boundary layer flows. That is why the hybrid mesh with brick mesh type is being developed to handle the difficulties.

Terry February 28, 2001 05:46

Re: Using turbulence model for steady state analys
 
So in this case, the results for lift and drag won't be accurate using the triangular mesh. Can using different type of models and mesh density improve on the results?.

Thanks.

John C. Chien February 28, 2001 08:03

Re: Using turbulence model for steady state analys
 
(1). Most people use the brick type mesh. (2). This could be H-type mesh, O-type mesh, or combination of the two. (3). The hybrid type also could be used, with brick type around the wall and tri/tet mesh away from the wall. But sometimes, it is not easy to generate good hybrid mesh. (4). I would suggest that, you try the flow over a flat plate first to see how the mesh affect the solution. Once you have a good feeling about the mesh effect, you can try it on the airfoil.


All times are GMT -4. The time now is 09:35.